3D Facial Expression Recognition on BU-3DFE (train)
98.25AccuracyPointNet++ (Baseline)
Evaluation Results
| Method | Links | |
|---|---|---|
| PointNet++ (Baseline)Used part of BU-3DFE (%)=80, Point cloud configuration (Full vs 512 points)=5122026.03 | 98.25 | |
| PointNet++ (Baseline)Used part of BU-3DFE (%)=80, Point cloud configuration (Full vs 512 points)=Full2026.03 | 96.5 | |
| pointcloud refinement pipelineUsed part of BU-3DFE (%)=25, Point cloud configuration (Full vs 512 points)=5122026.03 | 89.31 | |
| pointcloud refinement pipelineUsed part of BU-3DFE (%)=10, Point cloud configuration (Full vs 512 points)=5122026.03 | 88.54 | |
| PointNet++ (Baseline)Used part of BU-3DFE (%)=10, Point cloud configuration (Full vs 512 points)=Full2026.03 | 84.38 | |
| FLAME-based 3D pointcloud generation pipelineUsed part of BU-3DFE (%)=25, Point cloud configuration (Full vs 512 points)=Full2026.03 | 78.42 | |
| PointNet++ (Baseline)Used part of BU-3DFE (%)=10, Point cloud configuration (Full vs 512 points)=5122026.03 | 72.92 | |
| PointNet++ (Baseline)Used part of BU-3DFE (%)=25, Point cloud configuration (Full vs 512 points)=Full2026.03 | 67.34 | |
| PointNet++ (Baseline)Used part of BU-3DFE (%)=25, Point cloud configuration (Full vs 512 points)=5122026.03 | 66.13 |